A micro-sized vaccine based on recombinant Lactiplantibacillus plantarum fights against SARS-CoV-2 infection via intranasal immunization

Abstract

COVID-19 has globally spread to burden the medical system. Even with a massive vaccination, a mucosal vaccine offering more comprehensive and convenient protection is imminent. Here, a micro-sized vaccine based on recombinant Lactiplantibacillus plantarum (rLP) displaying spike or receptor-binding domain (RBD) was characterized as microparticles, and its safety and protective effects against SARS-CoV-2 were evaluated. We found a 66.7% mortality reduction and 100% protection with rLP against SARS-CoV-2 in a mouse model. The histological analysis showed decreased hemorrhage symptoms and increased leukocyte infiltration in the lung. Especially, rLP:RBD significantly decreased pulmonary viral loads. For the first time, our study provides a L. plantarum-vectored vaccine to prevent COVID-19 progress and transmission via intranasal vaccination.

Graphical abstract

Respiratory mucosal defense against SARS-CoV-2 infection where recombinant Lactiplantibacillus plantarum displaying spike or receptor-binding domain (RBD) inhibits viral replication in the lung and prevents disease progress by intranasal inoculation.

1. Introduction

The coronavirus disease 2019 (COVID-19) has spread globally for over three years without signs of termination. As its infectious agent, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) causes amounts of clinical manifestations initially targeting the respiratory tract mucosa. The mortality of this disease is positively correlated with viral loads and worse lung tissue damage^1,2. Neutralizing the pulmonary virus seems an efficient strategy to prevent disease transmission and progression. There are few licensed inhalable or intranasal vaccines for COVID-19 because of the mucosal complexity and proper delivery system deficiency. Fortunately, licensed mucosal vaccines have fortified the frontiers against polioviruses, influenza viruses, and rotaviruses³. Mucosal vaccines targeting COVID-19 could be a new concern with many advantages4, 5, 6.

Unlike injectable vaccines, it is much easier for mucosal vaccines to provide a natural barrier to vulnerable sites. Another application would be a potentially new choice in combination with injectable vaccines during the prime/boost immunization, furnishing multi-dimensional protection against infectious agents or variant viruses. Notably, non-invasive inoculation seems to be a solution to alleviate or eliminate the antibody-dependent enhancement of COVID-19^4,7,8. The nano-microparticle technology will be potentially applied in the next-generation vaccine development due to its excellent distribution and immuno-stimulator characteristics.

Our previous studies^9,10 have constructed two recombinant Lactiplantibacillus plantarum (rLP) displaying the SARS-CoV-2 spike or receptor-binding domain (RBD) on the surface. Because of broader research and facilitation, L. plantarum as a micro-sized vector characterized by safety and tolerability can be a vehicle to deliver plasmids, enzymes, and proteins to the mucosal surface. Their strong adhesivity assists in mucosal colonization without disruption of normal flora¹¹. We have discovered that rLP with antigenic RBD intranasally administrated into BALB/c mice significantly induces specific secretory IgA in the lung and fecal besides cellular immunity¹⁰. Therefore, in this study, we used a virulent mouse-adapted SARS-CoV-2 variant strain, C57MA14, to challenge the mice intranasally vaccinated with rLP. The survival rate, growth curves, and a modified European Efficiency Production Index (EEI) were proposed to describe the comprehensive mitigation effects of intranasal vaccination on virus challenges. Using RT-qPCR and pathological analysis, we reveal the suppression of nasal prevention on the pulmonary viral titers and tissue inflammation.

2. Materials and methods

2.1. Animal and virus

Six-week-old female BALB/c mice (Vital River Laboratory Animal Technology, Beijing, China) were housed under pathogen-free standard environmental conditions (12 h light/dark cycle and 22–25 °C, 45%–50% relative humidity) and provided with standard food and water ad libitum. The experimental animal procedures were approved by the Laboratory Animal Welfare and Ethics Committee of the Changchun Veterinary Research Institute. The mouse-adapted SARS-CoV-2 strain C57MA14 was kindly provided by Prof. Gao¹².

2.2. Animal immunization

The recombinant L. plantarum (rLP) expressing SARS-CoV-2 spike (rLP:S) or RBD (rLP:RBD) were constructed in our previous works, and so did the rLP contained blank plasmid (LP:Vector)^9,10. Before vaccination, the LP:Vector, rLP:S, and rLP:RBD were cultured in the MRS broth (Hopebiol, Qingdao, China) containing 10 μg/mL erythromycin (Sigma–Aldrich, MO, USA) respectively. When the OD₆₀₀ of the broth reached 0.3–0.5, the inducer SppIP¹³ (GenScript Biotech, Nanjing, China) was added into the culture at a 100 ng/mL concentration. All the cultures were incubated at 37 °C for 6 h after the addition of the inducer. The bacterial pellets were collected by centrifuge and suspended in sterile PBS for vaccination. A total of 22 BALB/c mice were randomly divided into four groups, PBS (n = 6), LP:Vector (n = 5), rLP:S (n = 5), and rLP:RBD (n = 6) respectively. Every mouse was administered 20 μL of PBS (negative control) or PBS containing 1 × 10⁹ CFU of rLPs for three consecutive days (Days 1, 2, and 3), and booster immunizations were administered in the same way 14 days later (Days 15, 16, and 17).

2.3. Animal challenge

All the vaccinated mice were transferred into the ABSL-3 on the Day 28 post-prime immunizations and challenged with C57MA14. Briefly, every mouse was performed anesthesia by inhalation of isoflurane in turn. Then the 50 μL of fluid containing 5 × 10^3.5 median tissue culture infective dose (TCID₅₀) of SARS-CoV-2 was administrated by nasal drops. Since the day performed the challenge, the mouse was observed and weighed daily. At the 5 days and 7 days post-infection, 3 mice per group were euthanized, and their lungs and nasal cavity were collected. The lung was divided into two parts for evaluating viral loads and histopathological lesions.

The dead events were recorded daily, and the survival ratio was calculated with the Kaplan-Meier estimate. Each weight data (g), including the initial weight (g), was daily recorded and calculated with Eq. (1):

Initial weight (%) = (Real weight /Initial weight) × 100 (1)

where Initial weight (%) is the percentage of daily recorded weight to initial weight indicating the weight change, Real weight is the daily recorded weight data of each mouse, and Initial weight is firstly recorded weight data of each mouse.

The mean weight and initial weight percent curves were drawn with GraphPad Prism (GraphPad Software, CA, USA). Error bars indicate standard deviation.

2.4. Relative efficiency production index (REPI) creation

Herein, we modified European Efficiency Production Index, usually utilized in evaluating flock growth performance¹⁴, as the relative European Efficiency Production Index (REPI). Feed conversion ratios remain stable due to the same feeding environment and additives. REPI was calculated by Eq. (2):

$$\text{REPI}=\frac{\text{Liveability}\left(\%,\text{vaccination}\right)\times \text{Body}\text{weight}\left(\mathrm{g}\right)}{\text{Liveability}\left(\%,\text{challenge}\right)\times \text{Body}\text{weight}\left(\mathrm{g},\text{challenge}\right)}$$ (2)

where Livability is the survival ratio of infected mice in vaccination or viral challenge group at 5 days post-infection, Body weight (g) is the real weight data of each mouse in vaccination or viral challenge groups at 5 days post-infection, and the Body weight (g, challenge) is the average body weight grams in the viral challenge group at 5 days post-infection.

2.5. Real-time PCR (RT-qPCR)

For detecting the viral loads in the main organs of infected mice, the tissues of lungs, nasal cavity, spleens, and intestines were homogenized respectively in PBS buffer as a ratio of 0.1 g tissue in 0.5 mL PBS. Total RNA was extracted from the 100 μL homogenate using TIANamp Virus RNA Kit (TIANGEN Biotech, Beijing, China). The viral loads in tissues were quantified using HiScript II U⁺ One Step qRT-PCR Probe Kit (Vazyme Biotech, Nanjing, China) with primers targeting the N gene. The sequences of primers were: 5′-GGGGAACTTCTCCTGCTAGAAT-3′ (forward), 5′-CAGACATTTTGCTCTCAAGCT-3′ (reverse), and 5′-FAM-TTGCTGCTGCTTGACAGATT-6-carboxytetramethylrhodamine (TAMRA)-3′ (TaqMan probe)¹⁵.

2.6. Histopathological examination

The part of the lung or nasal cavity tissue was fixed in 4% paraformaldehyde solutions for 7 days, paraffin-embedded, sectioned, and stained with hematoxylin and eosin (H&E) according to standard protocols. These sections were imaged using light microscopy and analyzed as described previously¹⁶. The pathological changes such as infiltration, congestion, epithelial cell death, widened alveolar septum, inflammatory cell infiltration, and collapse of the alveolar structure in the lungs were characterized.

2.7. Western blot

The expressions of SARS-CoV-2 spike or RBD in the induced rLPs were verified by Western blot before immunization. In brief, all the microparticles were crushed with 0.1 μm glass beads (Sigma–Aldrich) for 20 min, contributing to bacterial protein in lysates. The lysates were mixed with 5× loading buffer (Beyotime Biotechnology, Shanghai, China) and boiled in a water bath for complete denaturation. The protein samples were separated by 10% SDS-PAGE gel and electro-transferred onto a nitrocellulose membrane (Whatman, Maidstone, UK). The membrane was blocked by 5% skimmed milk solution at room temperature for 2 h, followed by incubating with anti-HA tag rabbit polyclonal antibody (1/1000, v/v, Proteintech China, Wuhan, China) at room temperature for 2 h and washed with TBST four times. Afterward, the membrane was incubated with horseradish peroxidase (HRP)-conjugated secondary antibody (1/10,000, v/v, Beyotime Biotechnology) at room temperature for 1 h, and washed with TBST four times. Subsequently, the proteins were visualized using enhanced chemiluminescence (ECL) reagent (Thermo Fisher Scientific, MA, USA) in GE Amersham Imager600 (General Electric Company, MA, USA).

2.8. Indirect immunofluorescence assay (IFA)

The induced LP:Vector, rLP:S, and rLP:RBD were harvested and washed twice with PBS, subsequently suspended and incubated with a primary antibody (anti-HA tag rabbit polyclonal antibody, 1/100, v/v, Proteintech China) at 4 °C overnight. Then, the samples were washed with PBS three times and incubated with a secondary antibody (FITC-conjugated goat anti-rabbit IgG, 1/5000, v/v, Zsbio Co., Ltd., Beijing, China) for 40 min at room temperature and washed again. The suspended LP:Vector or rLPs were fixed on a clean coverslip in the dark and imaged by fluorescence microscopy (ZEISS, Oberkochen, Germany).

2.9. Flow cytometric analysis (FCM)

The induced LP:Vector, rLP:S, and rLP:RBD were harvested and marked antibodies as prior described. The suspended LP:Vector or rLPs were examined using a CytoFLEX flow cytometer (Beckman Coulter, CA, USA) with a FITC tunnel.

2.10. Size distribution analysis of microparticles

The induced LP:Vector, rLP:S, and rLP:RBD were prepared. The bacterial pellets were collected by centrifuge, suspended fully, and diluted to OD₆₀₀ 0.3–0.5 in deionized water. The size distribution of microparticles was measured using a Nanoparticle Size Analyzer (Winner, Jinan, China) based on the principle of dynamic light scattering with a 90° of scattering angle and 532 nm of laser wavelength. Samples were added into quartz cuvettes with a 10-mm path length and equilibrated to a working temperature of 25 °C for 60 s before each measurement. Each sample was measured in 6-time after preparation. The mean z-average diameter and mean polydispersity index were characterized.

2.11. The size measurement of microparticles in transmission electron microscopy (TEM)

The induced L. plantarum, including LP18, LP:Vector, rLP:S, and rLP:RBD, were harvested and washed twice with deionized water. The microparticles were then negatively stained with 0.5% phosphotungstic acid (PTA, Sigma-Aldrich) and transferred onto 150-mesh formvar grids, subsequently observing and imaging under a transmission electron microscope Hitachi-7650 at an accelerating voltage of 90 kV. The standardized images of microparticles were analyzed by Photoshop software. Briefly, the pixel values of length (PL), width (PW), and area (PA) of microparticles in the images were measured using the ruler and lasso tools, and then the pixel values were converted into the real values (RL, RW, RA) according to the reference bar of 1 μm and its pixel value (PRB). The calibration Eqs. (3), (4), (5) were shown as follows:

RL (μm) = PL/PRB × 1 (3)

RW (μm) = PW/PRB × 1 (4)

RA (μm²) = PA/PRB² × 1 (5)

where RL/RW/RA is the calculated length/width/area value of the microparticles, PL is the pixel values of length, PW is the pixel values of width, PA is the pixel values of area, PRB is the pixel value of the reference bar of 1 μm.

2.12. Colony counting of LP:Vector and rLPs in the tissue of mice

The nasal cavity and lung of the mouse were collected respectively at 0.5, 3, 12, 24, 48,72, 120 h after PBS, LP:Vector, rLP:S, rLP:RBD intranasal administration, on the other hand, the brain and olfactory bulb of the mouse were collected respectively at 1, 3, 7 days post-inoculation. Then, 0.1 g of tissues were homogenized with 500 μL of PBS. The tissue homogenates were serially diluted 10 times in PBS, subsequently, diluted fluids were taken 5 μL and dropped onto erythromycin MRS plates. Every diluted fluid was performed with 2–3 duplicated dripping. After the plates were incubated at 37 °C for 24 h, the colonies were counted. Further, the numbers of lived rLPs in the nasal cavity and lung could be calculated by Eq. (6):

NLBT (CFU/g) = ANCC × Dilution ratio (10ⁿ) × 100/Tissue weight (0.1 g) (6)

where NLBT (CFU/g) is numbers of lived bacteria in tissue, ANCC is the average numbers of colony counts which are measured by counting the numbers of bacteria on the MRS plate, Dilution ratio may be any proper dilution from 10°–10⁶, 100 is the ratio of drop in tissue homogenates (500 μL/5 μL).

2.13. Cytotoxicity assay

To evaluate the cytotoxicity of the rLP:S and rLP:RBD, 1 × 10⁴ of pulmonary cells A549 and BEAS-2B were respectively seeded in 96-well plates and were co-cultured respectively with 10, 10³, 10⁵, and 10⁷ CFU/mL of the rLP:S or rLP:RBD in maintenance medium (containing 2% of fetal bovine serum) for 48 h. The cell viability was analyzed using Cell Counting Kit-8 (Dojindo Laboratories, kumamoto, Japan).

2.14. Blood routine and biochemical analysis

The count and classification of blood cells were tested using an automatic hematology analyzer (URIT, Guilin, China). About 100 μL of anticoagulant blood were sucked into an automated blood cell counter. The LDH in serum and BALF was tested by the automatic biochemical analyzer (URIT).

2.15. Pseudovirus-based neutralization assay

The SARS-CoV-2 strain WH-1 pseudovirus (DaRui biotech, Guangzhou, China) was used to assess the neutralizing antibody in BALF and fecal samples. Briefly, the BALF and fecal samples were sterilized by filtration using a 0.22 μm filter membrane. After inactivating the complement in a 56 °C water bath for 30 min, these samples were performed with an initial dilution of 1/2 (v/v) followed by 3-fold serial dilution in DMEM. The diluted samples were incubated with pseudovirus for 1 h at 37 °C, together with the virus control and cell control wells in every plate. Subsequently, 2 × 10⁴ freshly trypsinized HEK293T-ACE2 cells were added to each well and cultured for 48 h. Then, the culture supernatant was aspirated gently to leave 100 μL in each well and equilibrated to room temperature. 100 μL of luciferase substrate (Beyotime) was added to each well and incubated at room temperature for 2 min. 200 μL of lysate was transferred to white solid 96-well plates for the detection of luminescence using a microplate luminometer (Tecan, Mannedorf, Switzerland). The Tissue Culture Inhibiting Dilution neutralizing 50% of the infection (ID50) was calculated by nonlinear regression with Prism (GraphPad Software).

2.16. Statistical analysis

Statistical significance was determined by one-way or two-way analyses of variance (ANOVAs) with Bonferroni tests using GraphPad Prism (GraphPad Software). Statistical significance is shown as follows: ∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001.

3. Results and discussion

3.1. Characterization of rLP

Firstly, we detected the heterogeneous expression of rLP with an anti-HA tag or RBD antibody by Western blot. Certain binding proteins were visualized at the expected molecular weight (Supporting Information Fig. S1A and B). The confirmation from both antibodies proved a successful expression. To identify the size of LP vectors or rLPs, we utilized the particle size analyzer-based dynamic light scattering (DLS) principle to draw volume and cumulative distribution by multiple measures (Fig. 1A and B). The diameters were concentrated between 400 and 2700 nm, covering over 90% of the total. The polydispersity index from DLS for each group was almost in 0.1–0.4 (Fig. 1C), representing nearly monodisperse preparation¹⁷. Although we compared mean diameters data from DLS measurement and discovered a significant difference between each group (Fig. 1D), we still checked the size with the transmission electron microscope (TEM) to determine the particle size (Fig. 1E). Measured with image tools, each group's the length, width, and area showed no significant variation (Fig. 1F–H). The discrepancy between DLS and TEM seems to be due to a larger hydrodynamic diameter displaying the proteins on the rLP surface. In any case, we finally confirmed that rLP is a micron-sized delivery system.

Figure 1.

The size distribution profile of microparticles was measured by DLS and TEM. (A) Schematic illustrating the preparation of recombinant Lactiplantibacillus plantarum expressing SARS-CoV-2 antigen. The comparison between the size distribution of LP:Vector and rLP was shown with volume distribution (B) and cumulative size distribution (C). The average hydrodynamic diameters (D) and polydispersity indexes (E) of LP or rLPs were measured in the nanoparticle size analyzer. Data are presented as mean ± standard deviation (SD) (n ≥ 6); ∗P < 0.05 vs. indicated. ns, no significance. The TEM images of microparticles, including LP and two rLPs, were shown (F), and the average length (G), width (H), and area (I) of microparticles were analyzed by image measurement tools. Scale bar = 1 μm.

3.2. Safety evaluation

Unlike injectable vaccines, those by nasal administration need to be paid more attention to due to the nose-to-brain pathway. A recent study reported that SARS-CoV-2 RBD could damage the olfactory region and cause neurological toxicity¹⁸. As the route to the central nervous system (CNS), the olfactory axon in humans owns a diameter ranging from 0.1 to 0.7 μm¹⁹, and the average diameter is about 200 nm²⁰. Namely, particles less than 200 nm could be delivered into the CNS by intranasal administration. Our L. plantarum-vectored delivery system displays and anchors the recombinant protein on the surface (Fig. S1C and D), which is mostly larger than 700 nm. Therefore, these vaccines prefer entering the lung rather than invading the CNS to induce the immune response and provide safer protection. To confirm this assumption, we counted the potential live LP:Vector or rLPs in the brain and olfactory bulb from vaccinated mice. Unsurprisingly, we didn't find any bacterial colonies in both tissues which convinced us that L. plantarum does not enter into CNS by the nose-to-brain path (Supporting Information Fig. S2).

To learn their toxicity on respiratory tract mucosa, we added the LP:Vector and rLPs into the monolayer cultures with two kinds of pulmonary epithelial cells, BESA-2B and A549, as models. An outcome shows that low or medium doses of LP:vector and rLPs damaged no cells at 48 h post-inoculation, and even the high dose slightly affected the cell viability (Supporting Information Fig. S3A–C). Thus, we further analyzed the safety of rLPs in mice. Vaccinated intranasally with a normal dose of LP:Vector or rLPs, mice in each group suffered the routine blood analysis at 1, 3, and 7 days post-inoculation. The systemic check found white blood cells (WBCs), mainly neutrophils, increased beyond the normal range within 72 h post-inoculation. However, the complete recovery of WBCs parameters occurred within a week (Fig. 2). To explore whether the recruitment of neutrophils induced pulmonary epithelial cell death, we used lactate dehydrogenase (LDH) as an indicator of tissue damage. The level of LDH in both serum and bronchoalveolar lavage fluid (BALF) showed no difference between mock and other groups at any time post-inoculation (Supporting Information Fig. S4A and B). A continuous histological analysis indicated the tissues characterized with hemorrhage, inflammatory cell infiltration, etc. in lung tissue happened at the beginning of immunization. Then these gradually returned to normal status within a week corresponding to the results of routine blood analysis (Fig. 3A). According to the histological scores, we found that acute alveolar wall thickening and inflammatory cell infiltration in the mice post-inoculation are the main pathological features (Fig. 3B and C). The minor hemorrhage, congestion, and edema would not cause increased cell damage (Fig. 3D and E). The total scores of histology in each group show no significant difference (Fig. 3F). We discovered the structure of nasal mucosal epithelial cells was intact and normal according to the histological outcome (Fig. S5).

Figure 2.

Comparative analysis of white blood cells. The white blood cells (WBC) are analyzed by routine hematology at 1, 3, and 7 days post-inoculation. The number of WBC (A), lymphocytes (B), monocytes (C), neutrophils (D), eosinophils (H), and basophils (J) were measured. The responding percentage of lymphocytes (E), monocytes (F), neutrophils (G), eosinophils (I), and basophils (K) were shown. Data are presented as mean ± SD (n = 4); ∗P < 0.05 vs. indicated. ns, no significance.

Figure 3.

Histological analysis. The histological morphology was evaluated by hematoxylin and eosin staining of lung sections at 1, 3, and 7 days post-inoculation (A). Each image is representative of a group of 3 mice. Scale bar = 1000 μm (up) for 20× magnification or 100 μm (down) for 200× magnification. According to the grading system, histological changes were scored for alveolar wall thickening (B), inflammatory cell infiltration (C), hemorrhage (D), and edema (E) from 0 (normal) to 5 (severity). The comprehensive scores induced by LP:Vector or rLPs were calculated (F).

3.3. Viral challenge and protective efficacy

To evaluate the efficacy of intranasal administration with microparticles, we designed four groups, including vaccination with PBS, LP, rLP:S⁹, and rLP:RBD¹⁰ against viral challenge (Fig. 4A). On the 5th day post-infection, we found a 66.7% reduction in viral infection mortality of vaccination compared with non-vaccination (Fig. 4B). Even LP vector can increase the survival rate, shown as a potential alternative protective role on the viral infection besides previously described adjuvants for COVID-19 vaccines²¹. We also found that rLP expressing viral antigens further prevented the death from viral infections, which boosted specific IgA induced by rLP in the lung, as previously described¹⁰. Surprisingly, there are no significant differences between control and vaccination among the alive mice's mean weight or initial weight ratio (Fig. 4C and D). The reason needs to be further investigated. Only individual weights or survival rates can not represent the overall situation. From the analysis results, REPI is a fine tune and supplement to evaluate vaccine efficacy in vivo. It indicates that both LP vector and rLP reduced the infectious damage. Compared with the LP vector, rLP showed more noticeable protective effects (Fig. 4E). This work reveals that REPI could be a more sensitive and precise parameter to compare vaccines or antiviral drug efficacy.

Figure 4.

The protective study of rLP expressing SARS-CoV-2 spike or RBD via intranasal administration. (A) The scheme of vaccination and viral challenge. (B) Survival analysis calculated by Kaplan-Meier estimate among PBS, LP:Vector, rLP:S, and rLP:RBD groups. (C) and (D) Weight and percent of initial weight were daily recorded and calculated to draw the curve. (E) REPI calculated as equation (2) with data at 5 days p.i. (F, G) Real-time PCR was used to quantify the C57MA14 RNA in both lung and nasal cavities. Data are shown as mean ± SD, n = 3. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001 and ∗∗∗∗P < 0.0001 vs. indicated. ns, no significance. Three replicates were performed. Primers targeted the N gene to count viral copies or loads per gram. (H) Hematoxylin and eosin staining of lung sections from BALB/c mice after virus challenge or intranasal vaccination. Each image is representative of a group of 3 mice. Scale bar = 1000 μm (up) or 100 μm (down).

3.4. Viral loads and pathological analysis

Besides, viral loads determine viral transmission and disease progression capability. We detected the viral loads of the nasal cavity, lung, spleen, and intestine, representing a possible virus infectious route. Similar to the previous report of this animal model, there were higher viral loads in both the nasal cavity and lung tissue from each group (Fig. S6A–D). Unlike in the nasal cavity, the viral loads in the lung have been significantly reduced by intranasal administration with rLP:RBD (Fig. 4F and G). As mentioned previously, this reflects specific sIgA on virus neutralization. The most striking finding was the difference in viral load between the lungs and the nasal cavity. Although not expected, another study presented a similar result that viral RNA copies were significantly lower in the lung than in the nasal cavity after intranasal administration with commercial vaccines²². We additionally proved a specific neutralization activity of bronchoalveolar lavage fluid and fecal diluent from vaccinated mice to SARS-CoV-2 strain WH-1 pseudovirus, indicating the existence of functional sIgA against viral infection (Supporting Information Fig. S7).

Furthermore, infected pulmonary tissue had diffuse alveolar or acute lung damage (DAD) (Fig. 2H). There are amounts of eosinophilic material or bronchial epithelial cells shedding in the alveolar lumen. The hyaline membrane can be occasionally seen. When vaccinated with LP vector or rLP, hemorrhage symptoms in the lung were diminished, likely to weaken viral lethality. We also discovered that vaccinated mice have a clear alveolar morphology with slight alveolar hyperplasia or mild leukocyte infiltration. It indicates that rLP is beneficial in relieving lung damage caused by viral infection. To further explore the protective mechanism, we counted the number of rLPs adhering to the lung and nasal cavity (Supporting Information Fig. S8A). Intriguingly, they retain in the lung for a longer time than in the nasal cavity. The rLP:RBD seems more adhesive than rLP:S (Fig. S8B and C). These may explain the lower viral loads in the lung of rLP:RBD vaccinated mice.

4. Conclusions

In summary, our study reveals a microparticle platform based on rLP as a mucosal vaccine potential triggering anti-SARS-CoV-2 activity via reducing pulmonary injuries and viral loads. A 66.7% reduction in mortality and a 100% protection rate prove the efficacy of rLP vaccines. The novel index REPI facilitates a more sensitive and comprehensive evaluation in vivo of vaccines. We simultaneously evaluated the safety of rLP vaccines with nasal administration. Like most vaccines or adjuvants, rLP induced a proper inflammatory response to recruit immune cells at the beginning of immunization. Our works confirm an effective and safe intranasal vaccine strategy, provide an invaluable preclinical trial and insight on the mucosal vaccine for COVID-19, and shed light on the antiviral mechanism.

Author contributions

Chang Li, Maopeng Wang, and Ningyi Jin conceived and designed the study. Letian Li and Jiayi Hao performed experiments including the construction and vaccination of recombinant Lactiplantibacillus plantarum, Chang Li, Maopeng Wang, and Yuhang Jiang analyzed data. Guoqing Zhang and Jing Chen feed animals. Letian Li, Pengfei Hao and Yuwei Gao challenged the virus. Chang Li, Maopeng Wang, and Ningyi Jin wrote this paper.

Conflicts of interest

The authors declare no conflicts of interest.

Appendix A. Supplementary data

The following are the Supplementary data to this article.